Тестирование статистических гипотез с использованием р-значений и доверительных интервалов

The aim of the study was to compare statistical conclusions, based on confidence intervals (CI) and p-value by the example of assessment of efficacy of psychopharmacotherapy in two groups of schizophrenic patients according to the results of evaluation of PANSS indices in dynamics.Material and methods: а total of 46 patients with schizophrenia (27 subjects with shift-like course of the disease and 19 subjects with continuous type of paranoid schizophrenia), who were admitted to hospital in acute psychotic state, received amisulpride at the average dose of 765,5 mg/day. The evaluation of mental status was carried out every week using PANSS scale. As comparable indices of efficacy of treatment, change in the mean PANSS total score was used, which was calculated weekly each time of 9 visits. The hypothesis of statistical significance of differences in the use of Student’s t-test was checked by mathematical method, taking into account p-value and CI.Results and discussion: it was established, that during pharmacotherapy statistically significant differences of PANSS mean total score values between the two examined groups appeared since visit 6, on the 35th day of observation ( p ≤ 0,05). Since that very day 95% CI did not include zero value. This testifies to the fact, that mean values of the mean total score in the two studied groups in dynamics have statistically significant differences.Conclusions: the obtained results demonstrate consistency of statistical conclusions on the basis of p-values and 95% CI, which indicates appropriateness of adduction not only p-values, but first of all 95% CI in scientific publications.

шизофрения, PANSS, амисульприд, статистические гипотезы, t-критерий Стьюдента, 95% ДИ, schizophrenia, PANSS, amisulpride, statistical hypotheses, Student’s t-test, 95% CI

1. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals Updated December 2016. URL: www.icmje.org

2. Afifi A., Azen S. Statistical analysis. A computer oriented ap- proach. N.Y.: Academic Press, 1979. Recommendations for the Conduct, Reporting, Editing, and Pub- lication of Scholarly Work in Medical Journals Updated December 2016. URL: www.icmje.org Afifi A., Azen S. Statistical analysis. A computer oriented ap- proach. N.Y.: Academic Press, 1979

3. Боровков А.А. Математическая статистика. Москва, 1984. Borovkov A.A. Matematicheskaja statistika. Moskva, 1984. (In Russ.)

4. Kramer G. Mathematical methods of statistics. 1948. Kramer G. Mathematical methods of statistics. 1948

5. Scheffe G. Analysis Of Variance. 1959. Scheffe G. Analysis Of Variance, 1959

6. Welch B.L. The generalization of «student's» problem when several different population variances are involved. Biometrika. 1947;34:28-35. doi: 10.2307/2332510

7. Satterthwaite F.E. An approximate distribution of estimates of variance components. Biometrics Bulletin. 1946;2:110-114. doi: 10.2307/3002019